Convolutional Neural Networks

Project: Write an Algorithm for a Dog Identification App


In this notebook, some template code has already been provided for you, and you will need to implement additional functionality to successfully complete this project. You will not need to modify the included code beyond what is requested. Sections that begin with '(IMPLEMENTATION)' in the header indicate that the following block of code will require additional functionality which you must provide. Instructions will be provided for each section, and the specifics of the implementation are marked in the code block with a 'TODO' statement. Please be sure to read the instructions carefully!

Note: Once you have completed all of the code implementations, you need to finalize your work by exporting the Jupyter Notebook as an HTML document. Before exporting the notebook to html, all of the code cells need to have been run so that reviewers can see the final implementation and output. You can then export the notebook by using the menu above and navigating to File -> Download as -> HTML (.html). Include the finished document along with this notebook as your submission.

In addition to implementing code, there will be questions that you must answer which relate to the project and your implementation. Each section where you will answer a question is preceded by a 'Question X' header. Carefully read each question and provide thorough answers in the following text boxes that begin with 'Answer:'. Your project submission will be evaluated based on your answers to each of the questions and the implementation you provide.

Note: Code and Markdown cells can be executed using the Shift + Enter keyboard shortcut. Markdown cells can be edited by double-clicking the cell to enter edit mode.

The rubric contains optional "Stand Out Suggestions" for enhancing the project beyond the minimum requirements. If you decide to pursue the "Stand Out Suggestions", you should include the code in this Jupyter notebook.


Why We're Here

In this notebook, you will make the first steps towards developing an algorithm that could be used as part of a mobile or web app. At the end of this project, your code will accept any user-supplied image as input. If a dog is detected in the image, it will provide an estimate of the dog's breed. If a human is detected, it will provide an estimate of the dog breed that is most resembling. The image below displays potential sample output of your finished project (... but we expect that each student's algorithm will behave differently!).

Sample Dog Output

In this real-world setting, you will need to piece together a series of models to perform different tasks; for instance, the algorithm that detects humans in an image will be different from the CNN that infers dog breed. There are many points of possible failure, and no perfect algorithm exists. Your imperfect solution will nonetheless create a fun user experience!

The Road Ahead

We break the notebook into separate steps. Feel free to use the links below to navigate the notebook.

  • Step 0: Import Datasets
  • Step 1: Detect Humans
  • Step 2: Detect Dogs
  • Step 3: Create a CNN to Classify Dog Breeds (from Scratch)
  • Step 4: Create a CNN to Classify Dog Breeds (using Transfer Learning)
  • Step 5: Write your Algorithm
  • Step 6: Test Your Algorithm

Step 0: Import Datasets

Make sure that you've downloaded the required human and dog datasets:

Note: if you are using the Udacity workspace, you DO NOT need to re-download these - they can be found in the /data folder as noted in the cell below.

  • Download the dog dataset. Unzip the folder and place it in this project's home directory, at the location /dog_images.

  • Download the human dataset. Unzip the folder and place it in the home directory, at location /lfw.

Note: If you are using a Windows machine, you are encouraged to use 7zip to extract the folder.

In the code cell below, we save the file paths for both the human (LFW) dataset and dog dataset in the numpy arrays human_files and dog_files.

In [32]:
import numpy as np
from glob import glob

# load filenames for human and dog images
human_files = np.array(glob("/data/lfw/*/*"))
dog_files = np.array(glob("/data/dog_images/*/*/*"))

# print number of images in each dataset
print('There are %d total human images.' % len(human_files))
print('There are %d total dog images.' % len(dog_files))
There are 13233 total human images.
There are 8351 total dog images.

Step 1: Detect Humans

In this section, we use OpenCV's implementation of Haar feature-based cascade classifiers to detect human faces in images.

OpenCV provides many pre-trained face detectors, stored as XML files on github. We have downloaded one of these detectors and stored it in the haarcascades directory. In the next code cell, we demonstrate how to use this detector to find human faces in a sample image.

In [33]:
import cv2                
import matplotlib.pyplot as plt                        
%matplotlib inline                               

# extract pre-trained face detector
face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_alt.xml')

# load color (BGR) image
img = cv2.imread(human_files[0])
# convert BGR image to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

# find faces in image
faces = face_cascade.detectMultiScale(gray)

# print number of faces detected in the image
print('Number of faces detected:', len(faces))

# get bounding box for each detected face
for (x,y,w,h) in faces:
    # add bounding box to color image
    cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
    
# convert BGR image to RGB for plotting
cv_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

# display the image, along with bounding box
plt.imshow(cv_rgb)
plt.show()
Number of faces detected: 1

Before using any of the face detectors, it is standard procedure to convert the images to grayscale. The detectMultiScale function executes the classifier stored in face_cascade and takes the grayscale image as a parameter.

In the above code, faces is a numpy array of detected faces, where each row corresponds to a detected face. Each detected face is a 1D array with four entries that specifies the bounding box of the detected face. The first two entries in the array (extracted in the above code as x and y) specify the horizontal and vertical positions of the top left corner of the bounding box. The last two entries in the array (extracted here as w and h) specify the width and height of the box.

Write a Human Face Detector

We can use this procedure to write a function that returns True if a human face is detected in an image and False otherwise. This function, aptly named face_detector, takes a string-valued file path to an image as input and appears in the code block below.

In [51]:
# returns "True" if face is detected in image stored at img_path
def face_detector(img_path):
    gray = cv2.imread(img_path, 0)
    faces = face_cascade.detectMultiScale(gray)
    return len(faces) > 0

(IMPLEMENTATION) Assess the Human Face Detector

Question 1: Use the code cell below to test the performance of the face_detector function.

  • What percentage of the first 100 images in human_files have a detected human face?
  • What percentage of the first 100 images in dog_files have a detected human face?

Ideally, we would like 100% of human images with a detected face and 0% of dog images with a detected face. You will see that our algorithm falls short of this goal, but still gives acceptable performance. We extract the file paths for the first 100 images from each of the datasets and store them in the numpy arrays human_files_short and dog_files_short.

Answer:

98.00% human images correctly classified.

18.00% dog images misclassified.

In [77]:
from tqdm import tqdm

human_files_short = human_files[:100]
dog_files_short = dog_files[:100]

#-#-# Do NOT modify the code above this line. #-#-#

## TODO: Test the performance of the face_detector algorithm 
## on the images in human_files_short and dog_files_short.

detected_human_files = 0
detected_dog_files = 0

print("Going through human images...")
for human_file in tqdm(human_files_short):
    if face_detector(human_file):
        detected_human_files += 1
print("Finished detecting faces on human images.")

print("Going through dog images...")
for dog_file in tqdm(dog_files_short):
    if face_detector(dog_file):
        detected_dog_files += 1
print("Finished detecting faces on dog images.")

print("{:.2%} human images correctly classified.".format(detected_human_files/len(dog_files_short)))
print("{:.2%} dog images misclassified.".format(detected_dog_files/len(human_files_short)))
  2%|▏         | 2/100 [00:00<00:06, 14.40it/s]
Going through human images...
100%|██████████| 100/100 [00:07<00:00, 13.65it/s]
  0%|          | 0/100 [00:00<?, ?it/s]
Finished detecting faces on human images.
Going through dog images...
100%|██████████| 100/100 [01:20<00:00,  3.92it/s]
Finished detecting faces on dog images.
98.00% human images correctly classified.
18.00% dog images misclassified.

We suggest the face detector from OpenCV as a potential way to detect human images in your algorithm, but you are free to explore other approaches, especially approaches that make use of deep learning :). Please use the code cell below to design and test your own face detection algorithm. If you decide to pursue this optional task, report performance on human_files_short and dog_files_short.

In [36]:
### (Optional) 
### TODO: Test performance of anotherface detection algorithm.
### Feel free to use as many code cells as needed.

Step 2: Detect Dogs

In this section, we use a pre-trained model to detect dogs in images.

Obtain Pre-trained VGG-16 Model

The code cell below downloads the VGG-16 model, along with weights that have been trained on ImageNet, a very large, very popular dataset used for image classification and other vision tasks. ImageNet contains over 10 million URLs, each linking to an image containing an object from one of 1000 categories.

In [37]:
import torch
import torchvision.models as models

# define VGG16 model
VGG16 = models.vgg16(pretrained=True)

# check if CUDA is available
use_cuda = torch.cuda.is_available()

# move model to GPU if CUDA is available
if use_cuda:
    VGG16 = VGG16.cuda()
Downloading: "https://download.pytorch.org/models/vgg16-397923af.pth" to /root/.torch/models/vgg16-397923af.pth
100%|██████████| 553433881/553433881 [00:05<00:00, 98117771.37it/s] 

Given an image, this pre-trained VGG-16 model returns a prediction (derived from the 1000 possible categories in ImageNet) for the object that is contained in the image.

(IMPLEMENTATION) Making Predictions with a Pre-trained Model

In the next code cell, you will write a function that accepts a path to an image (such as 'dogImages/train/001.Affenpinscher/Affenpinscher_00001.jpg') as input and returns the index corresponding to the ImageNet class that is predicted by the pre-trained VGG-16 model. The output should always be an integer between 0 and 999, inclusive.

Before writing the function, make sure that you take the time to learn how to appropriately pre-process tensors for pre-trained models in the PyTorch documentation.

In [38]:
from PIL import Image
import torchvision.transforms as transforms
from torch.autograd import Variable
def VGG16_predict(img_path):
    '''
    Use pre-trained VGG-16 model to obtain index corresponding to 
    predicted ImageNet class for image at specified path
    
    Args:
        img_path: path to an image
        
    Returns:
        Index corresponding to VGG-16 model's prediction
    '''
    
    ## TODO: Complete the function.
    ## Load and pre-process an image from the given img_path
    ## Return the *index* of the predicted class for that image
    
    img = Image.open(img_path)

    # Define transforms
    transform_pipeline = transforms.Compose([transforms.Resize((224,224)),
                                         transforms.ToTensor(),
                                         transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                                              std=[0.229, 0.224, 0.225])])
    # Apply transforms
    img = transform_pipeline(img)
                            
    # Insert a new axis; the input is a mini-batch
    img = img.unsqueeze(0)  # Insert the new axis at index 0 i.e. in front of the other axes/dims. 

    # Pytorch expects a Variable type
    img = Variable(img)

    # If the model is on GPU, send img to GPU as well 
    if use_cuda:
        img = img.cuda()
    
    # Model is loaded as VGG16
    VGG16.eval()
    prediction = VGG16(img)  # returns a tensor of tuples (img, class_label)
    prediction = prediction.data.numpy().argmax() # return the class with highest probability
    return prediction # integer between 0 and 999

(IMPLEMENTATION) Write a Dog Detector

While looking at the dictionary, you will notice that the categories corresponding to dogs appear in an uninterrupted sequence and correspond to dictionary keys 151-268, inclusive, to include all categories from 'Chihuahua' to 'Mexican hairless'. Thus, in order to check to see if an image is predicted to contain a dog by the pre-trained VGG-16 model, we need only check if the pre-trained model predicts an index between 151 and 268 (inclusive).

Use these ideas to complete the dog_detector function below, which returns True if a dog is detected in an image (and False if not).

In [39]:
### returns "True" if a dog is detected in the image stored at img_path
def dog_detector(img_path):
    ## TODO: Complete the function.
    class_index = VGG16_predict(img_path)
    if class_index >= 151 and class_index <= 268:
        return True
    return False

(IMPLEMENTATION) Assess the Dog Detector

Question 2: Use the code cell below to test the performance of your dog_detector function.

  • What percentage of the images in human_files_short have a detected dog?
  • What percentage of the images in dog_files_short have a detected dog?

Answer:

0.00% images in human_files_short have a detected dog.

100.00% images in dog_files_short have a detected dog.

In [64]:
### TODO: Test the performance of the dog_detector function
### on the images in human_files_short and dog_files_short.

detected_human_files = 0
detected_dog_files = 0

print("Going through human images...")
for human_file in tqdm(human_files_short):
    if dog_detector(human_file):
        detected_human_files += 1
print("Finished detecting faces on human images.")

print("Going through dog images...")
for dog_file in tqdm(dog_files_short):
    if dog_detector(dog_file):
        detected_dog_files += 1
print("Finished detecting faces on dog images.")

print("{:.2%} human images contain a dog.".format(detected_human_files/len(dog_files_short)))
print("{:.2%} dog images contain a dog.".format(detected_dog_files/len(human_files_short)))
  0%|          | 0/100 [00:00<?, ?it/s]
Going through human images...
  1%|          | 1/100 [00:00<01:14,  1.33it/s]
  2%|▏         | 2/100 [00:01<01:11,  1.37it/s]
  3%|▎         | 3/100 [00:02<01:09,  1.40it/s]
  4%|▍         | 4/100 [00:02<01:08,  1.41it/s]
  5%|▌         | 5/100 [00:03<01:06,  1.42it/s]
  6%|▌         | 6/100 [00:04<01:05,  1.43it/s]
  7%|▋         | 7/100 [00:04<01:04,  1.44it/s]
  8%|▊         | 8/100 [00:05<01:03,  1.45it/s]
  9%|▉         | 9/100 [00:06<01:02,  1.45it/s]
100%|██████████| 100/100 [01:08<00:00,  1.45it/s]
  0%|          | 0/100 [00:00<?, ?it/s]
Finished detecting faces on human images.
Going through dog images...
100%|██████████| 100/100 [01:10<00:00,  1.45it/s]
Finished detecting faces on dog images.
0.00% human images contain a dog.
100.00% dog images contain a dog.

We suggest VGG-16 as a potential network to detect dog images in your algorithm, but you are free to explore other pre-trained networks (such as Inception-v3, ResNet-50, etc). Please use the code cell below to test other pre-trained PyTorch models. If you decide to pursue this optional task, report performance on human_files_short and dog_files_short.

In [ ]:
### (Optional) 
### TODO: Report the performance of another pre-trained network.
### Feel free to use as many code cells as needed.

Step 3: Create a CNN to Classify Dog Breeds (from Scratch)

Now that we have functions for detecting humans and dogs in images, we need a way to predict breed from images. In this step, you will create a CNN that classifies dog breeds. You must create your CNN from scratch (so, you can't use transfer learning yet!), and you must attain a test accuracy of at least 10%. In Step 4 of this notebook, you will have the opportunity to use transfer learning to create a CNN that attains greatly improved accuracy.

We mention that the task of assigning breed to dogs from images is considered exceptionally challenging. To see why, consider that even a human would have trouble distinguishing between a Brittany and a Welsh Springer Spaniel.

Brittany Welsh Springer Spaniel

It is not difficult to find other dog breed pairs with minimal inter-class variation (for instance, Curly-Coated Retrievers and American Water Spaniels).

Curly-Coated Retriever American Water Spaniel

Likewise, recall that labradors come in yellow, chocolate, and black. Your vision-based algorithm will have to conquer this high intra-class variation to determine how to classify all of these different shades as the same breed.

Yellow Labrador Chocolate Labrador Black Labrador

We also mention that random chance presents an exceptionally low bar: setting aside the fact that the classes are slightly imabalanced, a random guess will provide a correct answer roughly 1 in 133 times, which corresponds to an accuracy of less than 1%.

Remember that the practice is far ahead of the theory in deep learning. Experiment with many different architectures, and trust your intuition. And, of course, have fun!

(IMPLEMENTATION) Specify Data Loaders for the Dog Dataset

Use the code cell below to write three separate data loaders for the training, validation, and test datasets of dog images (located at dog_images/train, dog_images/valid, and dog_images/test, respectively). You may find this documentation on custom datasets to be a useful resource. If you are interested in augmenting your training and/or validation data, check out the wide variety of transforms!

In [7]:
import os
import torch
import torchvision.models as models

import numpy as np
from torchvision import datasets
from PIL import Image
import torchvision.transforms as transforms
from torch.autograd import Variable

### TODO: Write data loaders for training, validation, and test sets
## Specify appropriate transforms, and batch_sizes
from torchvision import datasets
import torchvision.transforms as transforms
from torch.utils.data.sampler import SubsetRandomSampler

# number of subprocesses to use for data loading
num_workers = 0
# how many samples per batch to load
batch_size = 20
# percentage of training set to use as validation
valid_size = 0.2

# convert data to a normalized torch.FloatTensor
train_transform = transforms.Compose([
    transforms.RandomHorizontalFlip(), # randomly flip and rotate
    transforms.RandomAffine(degrees=10, translate = (0.1, 0.1)),
    transforms.Resize((224,224)),
    transforms.ColorJitter(),
    transforms.ToTensor(),
    transforms.Normalize(mean=[0.485, 0.456, 0.406],
                         std=[0.229, 0.224, 0.225])
    ])

test_transform = transforms.Compose([
    transforms.Resize((224,224)),
    transforms.ToTensor(),
    transforms.Normalize(mean=[0.485, 0.456, 0.406],
                         std=[0.229, 0.224, 0.225])
    ])

# choose the training and test datasets
train_data = datasets.ImageFolder('/data/dog_images/train', transform=train_transform)
valid_data = datasets.ImageFolder('/data/dog_images/valid', transform=test_transform)
test_data = datasets.ImageFolder('/data/dog_images/test', transform=test_transform)

# print out some data stats
print('Num training images: ', len(train_data))
print('Num validation images: ', len(valid_data))
print('Num test images: ', len(test_data))


# define dataloader parameters
batch_size = 20
num_workers=0

# prepare data loaders
train_loader = torch.utils.data.DataLoader(train_data, batch_size=batch_size, 
                                           num_workers=num_workers, shuffle=True)
valid_loader = torch.utils.data.DataLoader(valid_data, batch_size=batch_size, 
                                          num_workers=num_workers, shuffle=True)
test_loader = torch.utils.data.DataLoader(test_data, batch_size=batch_size, 
                                          num_workers=num_workers, shuffle=True)
Num training images:  6680
Num validation images:  835
Num test images:  836

Question 3: Describe your chosen procedure for preprocessing the data.

  • How does your code resize the images (by cropping, stretching, etc)? What size did you pick for the input tensor, and why?
  • Did you decide to augment the dataset? If so, how (through translations, flips, rotations, etc)? If not, why not?

Answer:

The code resizes the images by stretching, scaling to size 224x224. I chose this size because model VGG16 accepts an input tensor of size 224.

I decided to augment the dataset through translations, rotations (function RandomAffine), flips and color alterations (ColorJitter). It should help the model generalize better.

In [9]:
from matplotlib import colors, cm, pyplot as plt
%matplotlib inline

from PIL import ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True

# helper for visualization
def imshow(img):
    img_min, img_max = img.min(), img.max()
    new_img = (img - img_min) / (img_max - img_min) 
    plt.imshow(np.transpose(new_img, (1,2,0)))

# obtain one batch of training images
dataiter = iter(train_loader)
images, labels = dataiter.next()
images = images.numpy() # convert images to numpy for display
classes = train_data.classes

# plot the images in the batch, along with the corresponding labels
fig = plt.figure(figsize=(20, 20))
for idx in np.arange(20):
    ax = fig.add_subplot(4, 20/4, idx+1, xticks=[], yticks=[])
    #plt.imshow(np.transpose(images[idx], (1, 2, 0)))
    imshow(images[idx])
    ax.set_title(classes[labels[idx]])

(IMPLEMENTATION) Model Architecture

Create a CNN to classify dog breed. Use the template in the code cell below.

In [88]:
num_classes
Out[88]:
133
In [120]:
import torch.nn as nn
import torch.nn.functional as F

num_classes = len(classes)

# define the CNN architecture
class Net(nn.Module):
    ### TODO: choose an architecture, and complete the class
    def __init__(self):
        super(Net, self).__init__()
        ## Define layers of a CNN
        self.conv_1 = nn.Conv2d(3, 16, 3, padding=1)
        self.pool = nn.MaxPool2d(2,2)
        # image size: 112 x 112 x 8
        self.conv_2 = nn.Conv2d(16, 32, 3, padding=1)
        # after pooling: 56 x 56 x 16
        self.conv_3 = nn.Conv2d(32, 64, 3, padding=1)
        # after pooling: 28 x 28 x 32
        #self.conv_4 = nn.Conv2d(64, 128, 3, padding=1)
        # after pooling: 14 x 14 x 128
        # self.fc_1 = nn.Linear(128 * 14 * 14, 500)
        
        self.fc_1 = nn.Linear(64 * 28 * 28, 500)
        self.fc_2 = nn.Linear(500, num_classes)
        
        self.dropout = nn.Dropout(0.25)
    
    def forward(self, x):
        ## Define forward behavior
        x = self.pool(F.relu(self.conv_1(x)))
        x = self.pool(F.relu(self.conv_2(x)))
        x = self.pool(F.relu(self.conv_3(x)))
        x = x.view(-1, 64 * 28 * 28)
        x = self.dropout(x)
        x = F.relu(self.fc_1(x))
        x = self.dropout(x)
        x = self.fc_2(x) # do not put sigmoid because of the cross entropy loss!
        return x

#-#-# You so NOT have to modify the code below this line. #-#-#

# instantiate the CNN
model_scratch = Net()

# move tensors to GPU if CUDA is available
if use_cuda:
    model_scratch.cuda()

Question 4: Outline the steps you took to get to your final CNN architecture and your reasoning at each step.

Answer:

Dog breed classification is a more complex problem than, for example, MNIST, so the CNN architecture requires more than two hidden layers. I'll try with three hidden convolutional layers at the beginning, followed by ReLU and max pooling, and two linear layers and add more layers if necessary. I'll also incorporate dropout in order to prevent overfitting.

Each convolutional layer has stride = 3 and padding = 1 which keeps the image's width and height the same. Therefore, each max pooling layer yields an image half the size of the input image. The input image is of shape 224 x 224 x 3. After three convolutional layers followed by max pooling, the output shape is 28 x 28 x 64. In order to feed that to the linear layers, the output of the last max pooling layer has to be flattened. There are two linear layers, one with weights of shape (28 28 64, 500) and the last one with (500, num_classes=133).

I chose cross entropy loss so I don't add (log-)softmax as the last layer.

(IMPLEMENTATION) Specify Loss Function and Optimizer

Use the next code cell to specify a loss function and optimizer. Save the chosen loss function as criterion_scratch, and the optimizer as optimizer_scratch below.

In [121]:
import torch.optim as optim

### TODO: select loss function
criterion_scratch = nn.CrossEntropyLoss()

### TODO: select optimizer
optimizer_scratch = optim.Adam(model_scratch.parameters(), lr = 0.001)
In [122]:
optimizer_scratch
Out[122]:
Adam (
Parameter Group 0
    amsgrad: False
    betas: (0.9, 0.999)
    eps: 1e-08
    lr: 0.001
    weight_decay: 0
)

(IMPLEMENTATION) Train and Validate the Model

Train and validate your model in the code cell below. Save the final model parameters at filepath 'model_scratch.pt'.

In [123]:
loaders_scratch = {'train': train_loader, 'valid': valid_loader, 'test': test_loader}
In [124]:
0.0001*6000
Out[124]:
0.6
In [125]:
def train(n_epochs, loaders, model, optimizer, criterion, use_cuda, save_path):
    """returns trained model"""
    # initialize tracker for minimum validation loss
    valid_loss_min = np.Inf 
    
    for epoch in range(1, n_epochs+1):
        # initialize variables to monitor training and validation loss
        train_loss = 0.0
        valid_loss = 0.0
        
        ###################
        # train the model #
        ###################
        model.train()
        for batch_idx, (data, target) in enumerate(loaders['train']):
            # move to GPU
            if use_cuda:
                data, target = data.cuda(), target.cuda()
            
            ## find the loss and update the model parameters accordingly
            ## record the average training loss, using something like
            ## train_loss = train_loss + ((1 / (batch_idx + 1)) * (loss.data - train_loss))
            
            # clear the gradients of all optimized variables
            optimizer.zero_grad()
            # forward pass
            output = model(data)
            # calculate the batch loss
            loss = criterion(output, target)
            # backward pass
            loss.backward()
            # perform a single optimization step (parameter update)
            optimizer.step()
            # update training loss
            train_loss += ((1 / (batch_idx + 1)) * (loss.data - train_loss))
            
            # Print train loss every 100 batches
            if batch_idx % 100 == 0:
                print('Epoch {}, batch {} loss: {:.6f}'.format(epoch, batch_idx + 1, train_loss))
                
        ######################    
        # validate the model #
        ######################
        model.eval()
        for batch_idx, (data, target) in enumerate(loaders['valid']):
            # move to GPU
            if use_cuda:
                data, target = data.cuda(), target.cuda()
            ## update the average validation loss
            # forward pass
            output = model(data)
            # calculate the batch loss
            loss = criterion(output, target)
            # update average validation loss 
            valid_loss += ((1 / (batch_idx + 1)) * (loss.data - valid_loss))
    
        # print training/validation statistics 
        print('Epoch: {} \tTraining Loss: {:.6f} \tValidation Loss: {:.6f}'.format(
            epoch, 
            train_loss,
            valid_loss
            ))
        
        ## TODO: save the model if validation loss has decreased
        # save model if validation loss has decreased
        if valid_loss <= valid_loss_min:
            print('Validation loss decreased ({:.6f} --> {:.6f}).  Saving model ...'.format(
            valid_loss_min,
            valid_loss))
            torch.save(model.state_dict(), save_path)
            valid_loss_min = valid_loss
    # return trained model
    return model


# train the model
model_scratch = train(10, loaders_scratch, model_scratch, optimizer_scratch, 
                      criterion_scratch, use_cuda, 'model_scratch.pt')

# load the model that got the best validation accuracy
model_scratch.load_state_dict(torch.load('model_scratch.pt'))
Epoch 1, batch 1 loss: 4.891535
Epoch 1, batch 101 loss: 4.906397
Epoch 1, batch 201 loss: 4.897249
Epoch 1, batch 301 loss: 4.883974
Epoch: 1 	Training Loss: 4.870491 	Validation Loss: 4.691835
Validation loss decreased (inf --> 4.691835).  Saving model ...
Epoch 2, batch 1 loss: 4.844355
Epoch 2, batch 101 loss: 4.643748
Epoch 2, batch 201 loss: 4.583063
Epoch 2, batch 301 loss: 4.553397
Epoch: 2 	Training Loss: 4.542130 	Validation Loss: 4.375618
Validation loss decreased (4.691835 --> 4.375618).  Saving model ...
Epoch 3, batch 1 loss: 4.487786
Epoch 3, batch 101 loss: 4.330626
Epoch 3, batch 201 loss: 4.298523
Epoch 3, batch 301 loss: 4.284485
Epoch: 3 	Training Loss: 4.280242 	Validation Loss: 4.136584
Validation loss decreased (4.375618 --> 4.136584).  Saving model ...
Epoch 4, batch 1 loss: 4.085675
Epoch 4, batch 101 loss: 4.114819
Epoch 4, batch 201 loss: 4.118540
Epoch 4, batch 301 loss: 4.118309
Epoch: 4 	Training Loss: 4.121001 	Validation Loss: 4.193085
Epoch 5, batch 1 loss: 4.296367
Epoch 5, batch 101 loss: 4.023682
Epoch 5, batch 201 loss: 4.012903
Epoch 5, batch 301 loss: 4.009225
Epoch: 5 	Training Loss: 4.006977 	Validation Loss: 4.018776
Validation loss decreased (4.136584 --> 4.018776).  Saving model ...
Epoch 6, batch 1 loss: 3.309098
Epoch 6, batch 101 loss: 3.861412
Epoch 6, batch 201 loss: 3.903109
Epoch 6, batch 301 loss: 3.907658
Epoch: 6 	Training Loss: 3.902192 	Validation Loss: 3.958883
Validation loss decreased (4.018776 --> 3.958883).  Saving model ...
Epoch 7, batch 1 loss: 3.536313
Epoch 7, batch 101 loss: 3.776831
Epoch 7, batch 201 loss: 3.789355
Epoch 7, batch 301 loss: 3.789986
Epoch: 7 	Training Loss: 3.793351 	Validation Loss: 3.858575
Validation loss decreased (3.958883 --> 3.858575).  Saving model ...
Epoch 8, batch 1 loss: 3.545842
Epoch 8, batch 101 loss: 3.636596
Epoch 8, batch 201 loss: 3.651950
Epoch 8, batch 301 loss: 3.666533
Epoch: 8 	Training Loss: 3.674795 	Validation Loss: 3.762506
Validation loss decreased (3.858575 --> 3.762506).  Saving model ...
Epoch 9, batch 1 loss: 3.228273
Epoch 9, batch 101 loss: 3.547696
Epoch 9, batch 201 loss: 3.556422
Epoch 9, batch 301 loss: 3.569172
Epoch: 9 	Training Loss: 3.573636 	Validation Loss: 3.793169
Epoch 10, batch 1 loss: 3.202740
Epoch 10, batch 101 loss: 3.470764
Epoch 10, batch 201 loss: 3.503878
Epoch 10, batch 301 loss: 3.474912
Epoch: 10 	Training Loss: 3.473861 	Validation Loss: 3.753939
Validation loss decreased (3.762506 --> 3.753939).  Saving model ...

(IMPLEMENTATION) Test the Model

Try out your model on the test dataset of dog images. Use the code cell below to calculate and print the test loss and accuracy. Ensure that your test accuracy is greater than 10%.

In [126]:
def test(loaders, model, criterion, use_cuda):

    # monitor test loss and accuracy
    test_loss = 0.
    correct = 0.
    total = 0.

    model.eval()
    for batch_idx, (data, target) in enumerate(loaders['test']):
        # move to GPU
        if use_cuda:
            data, target = data.cuda(), target.cuda()
        # forward pass: compute predicted outputs by passing inputs to the model
        output = model(data)
        # calculate the loss
        loss = criterion(output, target)
        # update average test loss 
        test_loss = test_loss + ((1 / (batch_idx + 1)) * (loss.data - test_loss))
        # convert output probabilities to predicted class
        pred = output.data.max(1, keepdim=True)[1]
        # compare predictions to true label
        correct += np.sum(np.squeeze(pred.eq(target.data.view_as(pred))).cpu().numpy())
        total += data.size(0)
            
    print('Test Loss: {:.6f}\n'.format(test_loss))

    print('\nTest Accuracy: %2d%% (%2d/%2d)' % (
        100. * correct / total, correct, total))

# call test function    
test(loaders_scratch, model_scratch, criterion_scratch, use_cuda)
Test Loss: 3.730656


Test Accuracy: 13% (111/836)

Step 4: Create a CNN to Classify Dog Breeds (using Transfer Learning)

You will now use transfer learning to create a CNN that can identify dog breed from images. Your CNN must attain at least 60% accuracy on the test set.

(IMPLEMENTATION) Specify Data Loaders for the Dog Dataset

Use the code cell below to write three separate data loaders for the training, validation, and test datasets of dog images (located at dogImages/train, dogImages/valid, and dogImages/test, respectively).

If you like, you are welcome to use the same data loaders from the previous step, when you created a CNN from scratch.

In [1]:
## TODO: Specify data loaders
import os
import torch
import torchvision.models as models

import numpy as np
from torchvision import datasets
from PIL import Image, ImageFile
import torchvision.transforms as transforms
from torch.autograd import Variable

### TODO: Write data loaders for training, validation, and test sets
## Specify appropriate transforms, and batch_sizes

ImageFile.LOAD_TRUNCATED_IMAGES = True

# number of subprocesses to use for data loading
num_workers = 0
# how many samples per batch to load
batch_size = 20
# percentage of training set to use as validation
valid_size = 0.2

# convert data to a normalized torch.FloatTensor
train_transform = transforms.Compose([
    transforms.RandomHorizontalFlip(), # randomly flip and rotate
    transforms.RandomAffine(degrees=10, translate = (0.1, 0.1)),
    transforms.Resize((224,224)),
    transforms.ColorJitter(),
    transforms.ToTensor(),
    transforms.Normalize(mean=[0.485, 0.456, 0.406],
                         std=[0.229, 0.224, 0.225])
    ])

test_transform = transforms.Compose([
    transforms.Resize((224,224)),
    transforms.ToTensor(),
    transforms.Normalize(mean=[0.485, 0.456, 0.406],
                         std=[0.229, 0.224, 0.225])
    ])

# choose the training and test datasets
train_data = datasets.ImageFolder('/data/dog_images/train', transform=train_transform)
valid_data = datasets.ImageFolder('/data/dog_images/valid', transform=test_transform)
test_data = datasets.ImageFolder('/data/dog_images/test', transform=test_transform)

# print out some data stats
print('Num training images: ', len(train_data))
print('Num validation images: ', len(valid_data))
print('Num test images: ', len(test_data))


# define dataloader parameters
batch_size = 20
num_workers=0

# prepare data loaders
train_loader = torch.utils.data.DataLoader(train_data, batch_size=batch_size, 
                                           num_workers=num_workers, shuffle=True)
valid_loader = torch.utils.data.DataLoader(valid_data, batch_size=batch_size, 
                                          num_workers=num_workers, shuffle=True)
test_loader = torch.utils.data.DataLoader(test_data, batch_size=batch_size, 
                                          num_workers=num_workers, shuffle=True)

loaders_transfer = {
    'train': train_loader,
    'valid': valid_loader,
    'test': test_loader
}

num_classes = len(train_data.classes)
Num training images:  6680
Num validation images:  835
Num test images:  836

(IMPLEMENTATION) Model Architecture

Use transfer learning to create a CNN to classify dog breed. Use the code cell below, and save your initialized model as the variable model_transfer.

In [2]:
import torchvision.models as models
import torch.nn as nn

## TODO: Specify model architecture 

# Load the pretrained model from pytorch
model_transfer = models.resnet50(pretrained=True)

# Print out the model structure
print(model_transfer)

# Freeze feature parameters
for param_name, param in model_transfer.named_parameters():
    if "fc" not in param_name:
        param.requires_grad = False
        
# Replace the classification layer 
fc_in_features = model_transfer.fc.in_features
model_transfer.fc = nn.Linear(fc_in_features, num_classes)

# check if CUDA is available
use_cuda = torch.cuda.is_available()

if use_cuda:
    model_transfer = model_transfer.cuda()
Downloading: "https://download.pytorch.org/models/resnet50-19c8e357.pth" to /root/.torch/models/resnet50-19c8e357.pth
100%|██████████| 102502400/102502400 [00:04<00:00, 20871675.30it/s]
ResNet(
  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (relu): ReLU(inplace)
  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
  (layer1): Sequential(
    (0): Bottleneck(
      (conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (downsample): Sequential(
        (0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): Bottleneck(
      (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (2): Bottleneck(
      (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
  )
  (layer2): Sequential(
    (0): Bottleneck(
      (conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (downsample): Sequential(
        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): Bottleneck(
      (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (2): Bottleneck(
      (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (3): Bottleneck(
      (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
  )
  (layer3): Sequential(
    (0): Bottleneck(
      (conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (downsample): Sequential(
        (0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): Bottleneck(
      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (2): Bottleneck(
      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (3): Bottleneck(
      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (4): Bottleneck(
      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (5): Bottleneck(
      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
  )
  (layer4): Sequential(
    (0): Bottleneck(
      (conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (downsample): Sequential(
        (0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): Bottleneck(
      (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (2): Bottleneck(
      (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
  )
  (avgpool): AvgPool2d(kernel_size=7, stride=1, padding=0)
  (fc): Linear(in_features=2048, out_features=1000, bias=True)
)

Question 5: Outline the steps you took to get to your final CNN architecture and your reasoning at each step. Describe why you think the architecture is suitable for the current problem.

Answer:

After reading this paper, I chose ResNet50 as the base architecture. The dog dataset is rather small so I freeze the parameters from the layers from the feature extractor part. I replace the classification layer (model_transfer.fc) with one linear layer with 133 output nodes.

(IMPLEMENTATION) Specify Loss Function and Optimizer

Use the next code cell to specify a loss function and optimizer. Save the chosen loss function as criterion_transfer, and the optimizer as optimizer_transfer below.

In [6]:
import torch.optim as optim

criterion_transfer = nn.CrossEntropyLoss()
optimizer_transfer = optim.Adam(model_transfer.fc.parameters(), lr=0.001)

(IMPLEMENTATION) Train and Validate the Model

Train and validate your model in the code cell below. Save the final model parameters at filepath 'model_transfer.pt'.

In [8]:
def train(n_epochs, loaders, model, optimizer, criterion, use_cuda, save_path):
    """returns trained model"""
    # initialize tracker for minimum validation loss
    valid_loss_min = np.Inf 
    
    for epoch in range(1, n_epochs+1):
        # initialize variables to monitor training and validation loss
        train_loss = 0.0
        valid_loss = 0.0
        
        ###################
        # train the model #
        ###################
        model.train()
        for batch_idx, (data, target) in enumerate(loaders['train']):
            # move to GPU
            if use_cuda:
                data, target = data.cuda(), target.cuda()
            
            ## find the loss and update the model parameters accordingly
            ## record the average training loss, using something like
            ## train_loss = train_loss + ((1 / (batch_idx + 1)) * (loss.data - train_loss))
            
            # clear the gradients of all optimized variables
            optimizer.zero_grad()
            # forward pass
            output = model(data)
            # calculate the batch loss
            loss = criterion(output, target)
            # backward pass
            loss.backward()
            # perform a single optimization step (parameter update)
            optimizer.step()
            # update training loss
            train_loss += ((1 / (batch_idx + 1)) * (loss.data - train_loss))
            
            # Print train loss every 100 batches
            if batch_idx % 100 == 0:
                print('Epoch {}, batch {} loss: {:.6f}'.format(epoch, batch_idx + 1, train_loss))
                
        ######################    
        # validate the model #
        ######################
        model.eval()
        for batch_idx, (data, target) in enumerate(loaders['valid']):
            # move to GPU
            if use_cuda:
                data, target = data.cuda(), target.cuda()
            ## update the average validation loss
            # forward pass
            output = model(data)
            # calculate the batch loss
            loss = criterion(output, target)
            # update average validation loss 
            valid_loss += ((1 / (batch_idx + 1)) * (loss.data - valid_loss))
    
        # print training/validation statistics 
        print('Epoch: {} \tTraining Loss: {:.6f} \tValidation Loss: {:.6f}'.format(
            epoch, 
            train_loss,
            valid_loss
            ))
        
        ## TODO: save the model if validation loss has decreased
        # save model if validation loss has decreased
        if valid_loss <= valid_loss_min:
            print('Validation loss decreased ({:.6f} --> {:.6f}).  Saving model ...'.format(
            valid_loss_min,
            valid_loss))
            torch.save(model.state_dict(), save_path)
            valid_loss_min = valid_loss
    # return trained model
    return model

# train the model
n_epochs = 10
model_transfer = train(n_epochs, loaders_transfer, model_transfer, optimizer_transfer, criterion_transfer, use_cuda, 'model_transfer.pt')

# load the model that got the best validation accuracy (uncomment the line below)
model_transfer.load_state_dict(torch.load('model_transfer.pt'))
Epoch 1, batch 1 loss: 4.809375
Epoch 1, batch 101 loss: 3.678638
Epoch 1, batch 201 loss: 2.650561
Epoch 1, batch 301 loss: 2.160516
Epoch: 1 	Training Loss: 2.049367 	Validation Loss: 0.860703
Validation loss decreased (inf --> 0.860703).  Saving model ...
Epoch 2, batch 1 loss: 0.923257
Epoch 2, batch 101 loss: 0.774125
Epoch 2, batch 201 loss: 0.776646
Epoch 2, batch 301 loss: 0.780486
Epoch: 2 	Training Loss: 0.773528 	Validation Loss: 0.759498
Validation loss decreased (0.860703 --> 0.759498).  Saving model ...
Epoch 3, batch 1 loss: 0.975510
Epoch 3, batch 101 loss: 0.612596
Epoch 3, batch 201 loss: 0.601469
Epoch 3, batch 301 loss: 0.603822
Epoch: 3 	Training Loss: 0.605115 	Validation Loss: 0.800185
Epoch 4, batch 1 loss: 0.508096
Epoch 4, batch 101 loss: 0.522716
Epoch 4, batch 201 loss: 0.517670
Epoch 4, batch 301 loss: 0.518118
Epoch: 4 	Training Loss: 0.513669 	Validation Loss: 0.629138
Validation loss decreased (0.759498 --> 0.629138).  Saving model ...
Epoch 5, batch 1 loss: 0.195329
Epoch 5, batch 101 loss: 0.469291
Epoch 5, batch 201 loss: 0.442888
Epoch 5, batch 301 loss: 0.444809
Epoch: 5 	Training Loss: 0.449544 	Validation Loss: 0.660327
Epoch 6, batch 1 loss: 0.532655
Epoch 6, batch 101 loss: 0.398085
Epoch 6, batch 201 loss: 0.422068
Epoch 6, batch 301 loss: 0.425708
Epoch: 6 	Training Loss: 0.429246 	Validation Loss: 0.665824
Epoch 7, batch 1 loss: 0.266853
Epoch 7, batch 101 loss: 0.381359
Epoch 7, batch 201 loss: 0.349955
Epoch 7, batch 301 loss: 0.360809
Epoch: 7 	Training Loss: 0.367622 	Validation Loss: 0.601913
Validation loss decreased (0.629138 --> 0.601913).  Saving model ...
Epoch 8, batch 1 loss: 0.507535
Epoch 8, batch 101 loss: 0.343662
Epoch 8, batch 201 loss: 0.351474
Epoch 8, batch 301 loss: 0.365846
Epoch: 8 	Training Loss: 0.372945 	Validation Loss: 0.744079
Epoch 9, batch 1 loss: 0.274422
Epoch 9, batch 101 loss: 0.320836
Epoch 9, batch 201 loss: 0.323442
Epoch 9, batch 301 loss: 0.341552
Epoch: 9 	Training Loss: 0.344756 	Validation Loss: 0.665522
Epoch 10, batch 1 loss: 0.218013
Epoch 10, batch 101 loss: 0.298338
Epoch 10, batch 201 loss: 0.312107
Epoch 10, batch 301 loss: 0.315396
Epoch: 10 	Training Loss: 0.316647 	Validation Loss: 0.666817

(IMPLEMENTATION) Test the Model

Try out your model on the test dataset of dog images. Use the code cell below to calculate and print the test loss and accuracy. Ensure that your test accuracy is greater than 60%.

In [3]:
# load the model that got the best validation accuracy (uncomment the line below)
model_transfer.load_state_dict(torch.load('model_transfer.pt'))
In [4]:
model_transfer.cpu()
torch.save(model_transfer.state_dict(), "model_transfer.pt")
In [8]:
model_transfer.cuda()
Out[8]:
ResNet(
  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (relu): ReLU(inplace)
  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
  (layer1): Sequential(
    (0): Bottleneck(
      (conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (downsample): Sequential(
        (0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): Bottleneck(
      (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (2): Bottleneck(
      (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
  )
  (layer2): Sequential(
    (0): Bottleneck(
      (conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (downsample): Sequential(
        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): Bottleneck(
      (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (2): Bottleneck(
      (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (3): Bottleneck(
      (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
  )
  (layer3): Sequential(
    (0): Bottleneck(
      (conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (downsample): Sequential(
        (0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): Bottleneck(
      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (2): Bottleneck(
      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (3): Bottleneck(
      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (4): Bottleneck(
      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (5): Bottleneck(
      (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
  )
  (layer4): Sequential(
    (0): Bottleneck(
      (conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
      (downsample): Sequential(
        (0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): Bottleneck(
      (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
    (2): Bottleneck(
      (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)
      (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU(inplace)
    )
  )
  (avgpool): AvgPool2d(kernel_size=7, stride=1, padding=0)
  (fc): Linear(in_features=2048, out_features=133, bias=True)
)
In [9]:
def test(loaders, model, criterion, use_cuda):

    # monitor test loss and accuracy
    test_loss = 0.
    correct = 0.
    total = 0.

    model.eval()
    for batch_idx, (data, target) in enumerate(loaders['test']):
        # move to GPU
        if use_cuda:
            data, target = data.cuda(), target.cuda()
        # forward pass: compute predicted outputs by passing inputs to the model
        output = model(data)
        # calculate the loss
        loss = criterion(output, target)
        # update average test loss 
        test_loss = test_loss + ((1 / (batch_idx + 1)) * (loss.data - test_loss))
        # convert output probabilities to predicted class
        pred = output.data.max(1, keepdim=True)[1]
        # compare predictions to true label
        correct += np.sum(np.squeeze(pred.eq(target.data.view_as(pred))).cpu().numpy())
        total += data.size(0)
            
    print('Test Loss: {:.6f}\n'.format(test_loss))

    print('\nTest Accuracy: %2d%% (%2d/%2d)' % (
        100. * correct / total, correct, total))

test(loaders_transfer, model_transfer, criterion_transfer, use_cuda)
Test Loss: 0.657360


Test Accuracy: 80% (676/836)

(IMPLEMENTATION) Predict Dog Breed with the Model

Write a function that takes an image path as input and returns the dog breed (Affenpinscher, Afghan hound, etc) that is predicted by your model.

In [2]:
data_transfer = {'train': train_data, 'valid': valid_data, 'test': test_data}
In [3]:
import torch
import torchvision.models as models
import torch.nn as nn
import torch.nn.functional as F

## TODO: Specify model architecture 

# Load the pretrained model from pytorch
model_transfer = models.resnet50(pretrained=True)
num_classes = 133

# Replace the classification layer 
fc_in_features = model_transfer.fc.in_features
model_transfer.fc = nn.Linear(fc_in_features, num_classes)

model_transfer.load_state_dict(torch.load('model_transfer.pt'))
In [74]:
### TODO: Write a function that takes a path to an image as input
### and returns the dog breed that is predicted by the model.

# list of class names by index, i.e. a name can be accessed like class_names[0]
class_names = [item[4:].replace("_", " ") for item in data_transfer['train'].classes]

thresh = 0.3
min_thresh = 0.6

def predict_breed_transfer(img_path):
    # load the image and return the predicted breed
    img = Image.open(img_path)
        
    # Define transforms
    transform_pipeline = transforms.Compose([transforms.Resize((224,224)),
                                         transforms.ToTensor(),
                                         transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                                              std=[0.229, 0.224, 0.225])])
    # Apply transforms
    img = transform_pipeline(img)
                            
    # Insert a new axis; the input is a mini-batch
    img = img.unsqueeze(0)  # Insert the new axis at index 0 i.e. in front of the other axes/dims. 

    # Pytorch expects a Variable type
    img = Variable(img)
    
    # Model
    model_transfer.eval()
    
    with torch.no_grad():
        output = F.softmax(model_transfer(img), dim=1)  # returns a tensor of tuples (img, class_label)
        
        probs, indices = torch.sort(output, descending=True)
        probs, indices = probs[0], indices[0]
        
        if probs[0] > min_thresh:
            return class_names[indices[0]] + " ({:.2%})".format(probs[0])
        
        else:
            idxs = probs > thresh
            breed = class_names[indices[0]] + " ({:.2%})".format(probs[0])
            ind = 1
            while idxs[ind]:
                breed += ", " + class_names[indices[ind]] + " ({:.2%})".format(probs[ind])
                ind += 1

    return breed

Step 5: Write your Algorithm

Write an algorithm that accepts a file path to an image and first determines whether the image contains a human, dog, or neither. Then,

  • if a dog is detected in the image, return the predicted breed.
  • if a human is detected in the image, return the resembling dog breed.
  • if neither is detected in the image, provide output that indicates an error.

You are welcome to write your own functions for detecting humans and dogs in images, but feel free to use the face_detector and human_detector functions developed above. You are required to use your CNN from Step 4 to predict dog breed.

Some sample output for our algorithm is provided below, but feel free to design your own user experience!

Sample Human Output

(IMPLEMENTATION) Write your Algorithm

In [75]:
### TODO: Write your algorithm.
### Feel free to use as many code cells as needed.

import cv2
import cv2                
import matplotlib.pyplot as plt                        
%matplotlib inline       

face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_alt.xml')


def run_app(img_path):
    ## handle cases for a human face, dog, and neither
    img = cv2.imread(img_path)

    print("==============================================")
    if face_detector(img_path):
        breed = predict_breed_transfer(img_path)
        print("Hello human! You look like: {}".format(breed))
    elif dog_detector(img_path):
        breed = predict_breed_transfer(img_path)
        print("This dog looks like: {}".format(breed))
    else:
        print("Did not detect a dog or human in this image...")
        
    plt.imshow(Image.open(img_path))
    plt.show() # so the plot happens before print
            

Step 6: Test Your Algorithm

In this section, you will take your new algorithm for a spin! What kind of dog does the algorithm think that you look like? If you have a dog, does it predict your dog's breed accurately? If you have a cat, does it mistakenly think that your cat is a dog?

(IMPLEMENTATION) Test Your Algorithm on Sample Images!

Test your algorithm at least six images on your computer. Feel free to use any images you like. Use at least two human and two dog images.

Question 6: Is the output better than you expected :) ? Or worse :( ? Provide at least three possible points of improvement for your algorithm.

Answer: (Three possible points for improvement)

The output is as I expected.

Firstly, I could improve the face detector, maybe use one based on deep learning (using transfer learning) or try detecting face landmarks.

Secondly, dog detector needs improvement as well. In the output below, the VGG16 network recognized the fluffy chair as a dog. It could be also improved by transfer learning.

Lastly, by training the model for more epochs or/and increasing the model complexity by adding more layers, I could increase the accuracy of model_transfer.

In [78]:
## TODO: Execute your algorithm from Step 6 on
## at least 6 images on your computer.
## Feel free to use as many code cells as needed.

## suggested code, below
from glob import glob
import numpy as np

human_files = np.array(glob("/data/lfw/*/*"))
dog_files = np.array(glob("/data/dog_images/*/*/*"))
extra_files = np.array(os.listdir("./images"))
extra_files = [os.path.join("./images", extra_file) for extra_file in extra_files 
               if not extra_file.endswith("checkpoints") and "sample" not in extra_file and ".DS" not in extra_file]

for file in np.hstack((human_files[:3], dog_files[:3], extra_files)):
    run_app(file)
==============================================
Hello human! You look like: Havanese (19.07%)
==============================================
Hello human! You look like: American water spaniel (42.73%)
==============================================
Hello human! You look like: American water spaniel (36.65%)
==============================================
This dog looks like: Mastiff (49.53%)
==============================================
This dog looks like: Mastiff (48.10%)
==============================================
This dog looks like: Mastiff (35.88%)
==============================================
This dog looks like: Curly-coated retriever (46.71%), Boykin spaniel (34.12%)
==============================================
This dog looks like: Labrador retriever (97.34%)
==============================================
This dog looks like: Irish red and white setter (98.39%)
==============================================
This dog looks like: Labrador retriever (98.29%)
==============================================
This dog looks like: Mastiff (65.30%)
==============================================
This dog looks like: Labrador retriever (99.90%)
==============================================
Did not detect a dog or human in this image...
==============================================
This dog looks like: Curly-coated retriever (99.44%)
==============================================
This dog looks like: Labrador retriever (15.66%)
==============================================
This dog looks like: Cocker spaniel (64.91%)
==============================================
This dog looks like: Brittany (99.94%)
==============================================
This dog looks like: Icelandic sheepdog (49.02%)
In [ ]: